Emerging neurosurgical techniques offer potential breakthroughs in treatment of a growing spectrum of movement disorders and dementia (including Alzheimer?s disease, Tourette?s syndrome, autism, depression, and even obesity). These emerging surgical approaches extend the established success of deep-brain stimulation (DBS) in Parkinson?s disease by using novel electrode stimulators delivered trans-ventricularly to targets about the hypothalamus. While endoscopic approach provides reliable access to the ventricles, such access imparts a loss of cerebrospinal fluid (CSF) and brain shift up to ~10 mm in the very regions of interest for these novel DBS therapies. Therefore, realizing the benefit of such promising techniques requires advances beyond the state of the art in neuro-navigation. Moreover, the use of novel, directional electrodes in such techniques requires a means to guide and confirm stimulator placement. Especially in the early stages of development of such novel therapies, it is important to resolve uncertainties related to geometric precision in order to differentiate from underlying neurophysiology and other factors that may affect safety and outcome. We propose to develop and evaluate the following advances in intraoperative imaging, registration, and guidance to realize a platform for robot-assisted ventriculoscopic approach to deep-brain targets in a manner that overcomes conventional limitations of neuro-navigation and supports the emerging generation of novel DBS therapies: (Aim 1) Develop high-quality intraoperative cone-beam CT (CBCT) using 3D image reconstruction methods that propel image quality beyond conventional limits of CBCT, providing image quality sufficient to drive deformable registration with preop MRI, precisely localize stimulator placement, and provide a check against complication / intracranial hemorrhage. (Aim 2) Develop 3D-2D image registration methods to relate low-dose intraoperative radiographs with: (a) preop MRI for automatic patient registration; and (b) parametric models of DBS electrodes (including novel directional stimulators) for guidance and confirmation of stimulator placement with precision and accuracy beyond that of conventional tracking. (Aim 3) Develop multi-modality deformable image registration (MR-CBCT) to resolve alignment between preop MRI and intraoperative CBCT ? particularly peri-ventricular deep-brain deformation following CSF egress ? using a fast, modality- insensitive, diffeomorphic Demons method for accurate transformation of MRI / planning data to CBCT and endoscopy. (Aim 4) Develop endoscopic video registration to render 3D image and planning information directly in the endoscopic scene, providing accurate visualization of target and critical structures during ventriculoscopic approach. (Aim 5) Translate the methods from Aims 1-4 to clinical studies for quantitative evaluation of performance under realistic conditions, and combine within an integrated system for robot-assisted ventriculoscopy (RAV) approach to DBS targets. The proposal advances previous work in skull base and spinal neurosurgery, offering a high likelihood of success in enabling next-generation DBS, and creates an integrated system for image-guided surgical robotics beyond the state of the art ? a valuable testbed for development and translation of clinical systems under future academic-industry partnership.